Large-span arch bridge damage identification method and system
By constructing a bridge finite element model and internal product matrix diagram, combining transfer learning and noise condition processing, the problem of damage recognition of large span arch bridges under small samples is solved, and efficient and robust damage recognition effect is achieved, and the existing technology is overcome. The problem of insufficient recognition accuracy and noise immunity.
Patent Information
- Application Number
- CN202410957709.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-17
- Publication Date
- 2024-11-12
AI Technical Summary
The prior art is difficult to quickly and accurately perform damage recognition of large-span arch bridges under small samples, and it is difficult to achieve the expected recognition effect based on deep learning methods when the amount of data is limited.
By constructing a bridge finite element model, obtain structural acceleration data, establish an internal product matrix and convert it into an internal product matrix diagram, build a lightweight damage recognition model, and use transfer learning and fine-tuning training to obtain a large-span arch bridge damage recognition model, combined with noise conditions and plus Noise processing to improve recognition accuracy.
The full utilization of structural health monitoring data has been achieved, the efficiency and robustness of damage recognition of large-span arch bridges has been improved, the noise resistance has been greatly improved, and the identification problems under the influence of environmental interference and sensor instability have been solved.
Smart Images

Figure CN118940565A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of bridge structure health monitoring and damage assessment, and in particular to a damage identification method and system for a long-span arch bridge. Background Art
[0002] Large-span arch bridges have become an important type of bridge for transportation networks across mountains and canyons due to their significant advantages such as high structural rigidity, good economy, and strong adaptability to mountainous areas. However, with the increase in service life, the performance degradation of arch bridges is inevitable. Therefore, identifying the damage status of arch bridges and understanding their true service status are of great theoretical research significance and engineering practical value for ensuring the safe operation of the structure.
[0003] At present, arch bridge damage identification methods can be roughly divided into parameter-based damage identification methods, machine learning-based damage identification methods and deep learning-based damage identification methods. Among them, the parameter-based damage identification method mainly identifies the damage of the structure through the change of parameters such as mode. In theory, this method is not limited by the monitoring points. However, due to the many difficulties in measuring high-order modes in practical applications, most damages often mainly affect high-order modes, which limits the scope of use of this method.
[0004] The damage identification method based on machine learning mainly realizes damage identification by learning enough labeled samples in advance, so as to improve the recognition accuracy. This type of method can automatically learn and extract relevant information from input features, but its nonlinear modeling ability is relatively limited, it is difficult to capture the complex nonlinear relationship within the structure, and it cannot process big data well.
[0005] Compared with machine learning, deep learning can better extract features from big data, so damage identification methods based on deep learning have developed rapidly. However, this method usually requires the construction of a large-scale damage data set, which is a major challenge for real bridges with limited damage monitoring data. When the amount of data is limited, the commonly used deep learning neural networks often find it difficult to achieve the expected recognition effect.
[0006] In summary, how to quickly and accurately identify damage of long-span arch bridges with small samples has become an urgent problem that technicians in this field need to solve. Summary of the invention
[0007] In view of the shortcomings of existing methods and the needs of practical applications, in order to improve the damage identification ability of large-span arch bridges and solve the problem of rapid and accurate damage identification of large-span arch bridges under small sample conditions, the present invention provides a large-span arch bridge damage identification method, comprising the following steps:
[0008] A bridge finite element model is constructed based on a large-span arch bridge, and the bridge finite element model is used to obtain structural acceleration data under various damage conditions; an inner product matrix is established according to the structural acceleration data, and the inner product matrix is converted into an inner product matrix diagram; a lightweight damage identification model is constructed, and the lightweight damage identification model is preliminarily trained using conventional training data; the lightweight damage identification model after the preliminarily training is fine-tuned using the inner product matrix diagram to obtain a large-span arch bridge damage identification model; the actual response of the large-span arch bridge under operating conditions is obtained, and the large-span arch bridge damage identification model is used to perform damage identification according to the actual response. The present invention makes up for the deficiency of large-span arch bridges due to the scarcity of historical monitoring data through transfer learning, fully extracts the information in the scarce labeled data under small samples compared with large-scale damage data sets, realizes the full utilization of structural health monitoring data, and improves the efficiency and robustness of damage identification of large-span arch bridges. At the same time, the damage index represented by the inner product matrix fully mines the damage information of the bridge structure, greatly improves the strong anti-noise ability in the field of damage identification of large-span arch bridges, and also solves the problem that some vibration signals are changed under the influence of environmental interference, sensor instability and other factors, thereby affecting the detection results of bridge structure damage.
[0009] Optionally, the long-span arch bridge damage identification method further includes performing noise processing on the structural acceleration data according to noise conditions. The present invention simulates the actual service environment of the bridge by performing noise processing on the structural acceleration data, which is conducive to improving the accuracy of long-span arch bridge damage identification.
[0010] Optionally, the noise operating conditions include:
[0011] A first noise operating condition, the noise distribution of the first noise operating condition satisfies a normal distribution with a mean of 15 and a variance of 0; a second noise operating condition, the noise distribution of the second noise operating condition satisfies a normal distribution with a mean of 15 and a variance of 1.0; a third noise operating condition, the noise distribution of the third noise operating condition satisfies a normal distribution with a mean of 15 and a variance of 2.5; a fourth noise operating condition, the noise distribution of the fourth noise operating condition satisfies a normal distribution with a mean of 15 and a variance of 5.0.
[0012] Optionally, according to the second noise condition and the third noise condition, the structural acceleration data is subjected to noise processing using Latin hypercube sampling. The noise condition and noise processing method provided by the present invention meet the actual bridge service conditions and are further conducive to improving the accuracy of damage identification of large-span arch bridges.
[0013] Optionally, establishing an inner product matrix according to the structural acceleration data comprises the following steps:
[0014] According to the structural acceleration data, a time domain response of the structure is obtained; a reference response is set, and an inner product vector is obtained according to the reference response and the time domain response; and the inner product matrix is established through the inner product vector.
[0015] Optionally, the inner product vector satisfies the following formula:
[0016] R={R 1l (0),R 2l (0),…,R nl (0)}
[0017] Where R represents the time domain response x of the lth measurement point. l (t) is the inner product vector of the reference response, R nl (0) represents the value of the cross-correlation function between the time domain response of the nth measuring point and the time domain response of the lth measuring point when the time delay τ=0.
[0018] Optionally, the inner product matrix satisfies the following formula:
[0019]
[0020] R kl (0)= <x k (t),x l (t)>,k=1,2,…,n,l=1,2,…,n
[0021] Among them, M represents the inner product matrix, R kl (0) represents the time domain response x of the kth measurement point k (t) and the time domain response x of the lth measurement point l The value of the cross-correlation function of (t) when the time delay τ=0, <·> represents the inner product operator, and n represents the number of measuring points. The inner product matrix established by the present invention involves the time domain response of each measuring point and fully mines the damage information of the bridge structure.
[0022] Optionally, the conventional training data is an Image Net dataset. The present invention uses the Image Net dataset for preliminary training, so that the lightweight damage identification model has a feature extraction capability with strong generalization capability, which is conducive to improving the efficiency of damage identification of large-span arch bridges.
[0023] Optionally, the lightweight damage identification model is a Mobile ViT lightweight convolutional neural network. The lightweight damage identification model adopted by the present invention has low computational complexity, fast running speed, few network structure parameters, and small memory space occupation, further improving the damage identification speed of large-span arch bridges.
[0024] In the second aspect, in order to be able to efficiently execute the large-span arch bridge damage identification method provided by the present invention, the present invention also provides a large-span arch bridge damage identification system, including a processor, an input device, an output device and a memory, wherein the processor, the input device, the output device and the memory are interconnected, wherein the memory is used to store a computer program, the computer program includes program instructions, and the processor is configured to call the program instructions to execute the large-span arch bridge damage identification method as described in the first aspect of the present invention. The large-span arch bridge damage identification system of the present invention has a compact structure and stable performance, and can stably execute the large-span arch bridge damage identification method provided by the present invention, further improving the overall applicability and practical application capabilities of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] Figure 1 A flow chart of a method for identifying damage of a long-span arch bridge provided by an embodiment of the present invention;
[0026] Figure 2 A schematic diagram of a long-span arch bridge provided by an embodiment of the present invention;
[0027] Figure 3 A structural acceleration data diagram under various working conditions provided by an embodiment of the present invention;
[0028] Figure 4 A damage identification result diagram for a single damage case provided by an embodiment of the present invention;
[0029] Figure 5 A damage identification result diagram under multiple damage conditions provided by an embodiment of the present invention;
[0030] Figure 6 A framework diagram of a long-span arch bridge damage identification system provided by an embodiment of the present invention;
[0031] Figure 7 A schematic diagram of the structure of a large-span arch bridge damage identification device provided in an embodiment of the present invention. DETAILED DESCRIPTION
[0032] The specific embodiments of the present invention will be described in detail below. It should be noted that the embodiments described herein are only for illustration and are not intended to limit the present invention. In the following description, a large number of specific details are set forth in order to provide a thorough understanding of the present invention. However, it is obvious to those of ordinary skill in the art that these specific details do not need to be adopted to implement the present invention. In other examples, in order to avoid confusing the present invention, known circuits, software or methods are not specifically described.
[0033] Throughout the specification, references to "one embodiment," "an embodiment," "an example," or "an example" mean that a particular feature, structure, or characteristic described in conjunction with the embodiment or example is included in at least one embodiment of the present invention. Therefore, the phrases "in one embodiment," "in an embodiment," "an example," or "an example" appearing in various places throughout the specification do not necessarily all refer to the same embodiment or example. In addition, particular features, structures, or characteristics may be combined in one or more embodiments or examples in any suitable combination and / or subcombination. In addition, it should be understood by those of ordinary skill in the art that the figures provided herein are for illustrative purposes and that the figures are not necessarily drawn to scale.
[0034] See also Figure 1 , Figure 1 This is a flow chart of a method for identifying damage of a large-span arch bridge of the present invention. In order to improve the damage identification capability of a large-span arch bridge and solve the problem of quickly and accurately identifying damage of a large-span arch bridge under a small sample condition, the present invention provides a method for identifying damage of a large-span arch bridge, such as Figure 1 As shown, the method comprises the following steps:
[0035] S1. Construct a bridge finite element model based on a long-span arch bridge, and use the bridge finite element model to obtain structural acceleration data under various damage conditions.
[0036] See also Figure 2 In the embodiment, the long-span arch bridge meets the following conditions: the main arch axis of the long-span bridge adopts a catenary, the calculated span is 600m, the rise is 125m, the rise-span ratio f / l=1 / 4.8, the arch axis coefficient m=1.9, the single rib of the arch box adopts a single-box single-chamber variable height section, the arch top section is 8m high, the arch foot section is 12m high, and the width is 6.5m. The transverse center distance of the arch rib is 16.5m, and two ribs of the beam are arranged transversely, which is a parallel arch form. The concrete arch rib adopts a steel tube concrete truss structure as a rigid skeleton. The outer diameter of the upper chord tube of the truss is 900mm, and the wall thickness of the steel tube changes from 30mm to 35mm from the arch foot to the arch top. The outer diameter of the lower chord tube is 900mm, and the wall thickness changes from 35mm to 30mm from the arch foot to the arch top. The steel material is Q420qD25Z, the external concrete of the arch rib is C60, and the main chord tube is poured with C80 self-compacting micro-expansive concrete.
[0037] In this embodiment, the finite element model uses one of the Midas software, Abaqus software or Ansys Mechanical software, and according to the above-mentioned long-span arch bridge conditions, the finite element model of the bridge is constructed using the finite element model, and then as many damage conditions as possible are generated through the bridge finite element model, and the structural acceleration data under each condition is obtained at the same time, such as Figure 3 As shown, Figure 3 The horizontal axis represents the number of data points, the vertical axis represents the measurement points, and the vertical axis represents the acceleration response data.
[0038] Furthermore, after obtaining the structural acceleration data under various working conditions, the structural acceleration data is subjected to noise processing according to the noise working conditions.
[0039] Furthermore, the noise operating conditions include:
[0040] A first noise operating condition, wherein the noise distribution of the first noise operating condition satisfies a normal distribution with a mean of 15 and a variance of 0;
[0041] A second noise operating condition, wherein the noise distribution of the second noise operating condition satisfies a normal distribution with a mean of 15 and a variance of 1.0;
[0042] A third noise operating condition, wherein the noise distribution of the third noise operating condition satisfies a normal distribution with a mean of 15 and a variance of 2.5;
[0043] A fourth noise operating condition, wherein the noise distribution of the fourth noise operating condition satisfies a normal distribution with a mean of 15 and a variance of 5.0.
[0044] Specifically, according to the second noise condition and the third noise condition, the structural acceleration data is subjected to noise addition processing by using Latin hypercube sampling.
[0045] Please refer to Table 1, which is a table showing the recognition accuracy of the damage identification model for long-span arch bridges under different noise conditions.
[0046] Noise operating conditions Accuracy % Accuracy Recall F1 value The first noise condition 96.6 96.6% 96.6% 96.6% Second noise condition 94.9 95.0% 94.9% 95.0% The third noise condition 94.5 95.3% 94.5% 94.9% The fourth noise condition 93.6 94.0% 93.6% 93.8%
[0047] Table 1 Identification accuracy of damage identification model for long-span arch bridges under different noise conditions
[0048] According to Table 1, it can be seen that under various noise conditions, the damage identification model of large-span arch bridges has good identification effect, demonstrating the good noise resistance of the damage identification model of large-span arch bridges.
[0049] S2. Establish an inner product matrix according to the structural acceleration data, and convert the inner product matrix into an inner product matrix graph.
[0050] Step S2, establishing an inner product matrix according to the structural acceleration data, comprises the following steps:
[0051] According to the structural acceleration data, a time domain response of the structure is obtained; a reference response is set, and an inner product vector is obtained according to the reference response and the time domain response; and the inner product matrix is established through the inner product vector.
[0052] Specifically, the time domain response x1(t),x2(t),…,x n (t), take the response x of the measuring point l l(t) is the reference response, then the inner product vector is:
[0053] R={R 1l (0),R 2l (0),…,R nl (0)}
[0054] Where R represents the time domain response x of the lth measurement point. l (t) is the inner product vector of the reference response, R nl (0) represents the value of the cross-correlation function between the time domain response of the nth measuring point and the time domain response of the lth measuring point when the time delay τ=0.
[0055] Set the value of the reference point to the response of each measurement point, that is, l = 1, 2, ..., n, and you can get the inner product matrix:
[0056]
[0057] R kl (0)= <x k (t),x l (t)>,k=1,2,…,n,l=1,2,…,n
[0058] Among them, M represents the inner product matrix, R kl (0) represents the time domain response x of the kth measurement point k (t) and the time domain response x of the lth measurement point l (t) when the time delay τ = 0, <·> represents the inner product operator, and n represents the number of measurement points.
[0059] Further, converting the inner product matrix into an inner product matrix graph includes:
[0060] Plotting matrices: Use Matplotlib's imshow function or similar functions to plot matrices. You can also adjust the color map and color bar to enhance the readability of the graph.
[0061] Add labels: Add labels to the axes of the matrix so that you know which two vectors each element represents the inner product between.
[0062] S3. Construct a lightweight damage identification model, and use conventional training data to preliminarily train the lightweight damage identification model.
[0063] Specifically, a lightweight damage recognition model is constructed using the Mobile ViT lightweight convolutional neural network, and based on the Image Net dataset, the lightweight damage recognition model is preliminarily trained using a transfer learning strategy to transfer the feature extraction capability with strong generalization ability to the lightweight damage recognition model.
[0064] S4. Using the inner product matrix diagram, fine-tune the lightweight damage identification model after preliminary training to obtain a large-span arch bridge damage identification model.
[0065] Specifically, the inner product matrix diagrams are divided into a training set and a test set in a ratio of 8:2. In this embodiment, the training set of each working condition has only 200 inner product matrix diagrams, and the test set has 47 inner product matrix diagrams.
[0066] The inner product matrix diagram is used to fine-tune and verify the high-level parameters of the lightweight damage identification model after preliminary training, so as to obtain a damage identification model for a large-span arch bridge.
[0067] S5. Acquire the actual response of the long-span arch bridge under operating conditions, and use the damage identification model for the long-span arch bridge to perform damage identification according to the actual response.
[0068] See also Figure 4 and Figure 5 The horizontal axis represents the prediction results of the damage identification model for long-span arch bridges, and the vertical axis represents the actual results of the finite element model simulation. Figure 4 and Figure 5 It can be seen that the damage identification model for large-span arch bridges provided in this embodiment has high accuracy in both single damage identification and multi-damage identification. In addition, the damage identification model for large-span arch bridges only takes 2.2 seconds to identify an inner product matrix diagram, which is extremely fast and highly real-time.
[0069] See also Figure 6 In an embodiment, in order to efficiently execute the large-span arch bridge damage identification method provided by the present invention, the present invention also provides a large-span arch bridge damage identification system, including: an input device, an output device, a processor, and a memory, wherein the input device, the output device, the processor, and the memory are interconnected, and the memory contains program instructions, and the program instructions are used for the steps of the large-span arch bridge damage identification method. The large-span arch bridge damage identification system of the present invention has a compact structure and stable performance, and can stably execute the large-span arch bridge damage identification method of the present invention, further improving the overall applicability and practical application ability of the present invention.
[0070] In an embodiment, the processor may be a central processing unit (CPU), which may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc. An input device may be used to obtain data information. An output device may be used to output the results obtained by storing program instructions contained in a computer program in a memory provided by the present invention. The memory may include a read-only memory and a random access memory, and provide instructions and data to the processor. A portion of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the type of device.
[0071] In yet another alternative embodiment, see Figure 7 In order to efficiently execute the large-span arch bridge damage identification method provided by the present invention, this embodiment also provides a large-span arch bridge damage identification device, such as Figure 7 As shown, including:
[0072] The memory 10 is used to store computer programs; the processor 20 is used to execute the computer programs to implement the above-mentioned large-span arch bridge damage identification method.
[0073] Memory 10 , processor 20 , communication interface 31 and communication bus 32 . Memory 10 , processor 20 , communication interface 31 all communicate with each other via communication bus 32 .
[0074] In an embodiment, the memory 10 is used to store one or more program instructions. The memory 10 may store program instructions for implementing the following functions:
[0075] A bridge finite element model is constructed based on a large-span arch bridge, and the bridge finite element model is used to obtain structural acceleration data under various damage conditions; an inner product matrix is established according to the structural acceleration data, and the inner product matrix is converted into an inner product matrix diagram; a lightweight damage identification model is constructed, and the lightweight damage identification model is preliminarily trained using conventional training data; the lightweight damage identification model after the preliminarily training is fine-tuned using the inner product matrix diagram to obtain a large-span arch bridge damage identification model; the actual response of the large-span arch bridge under operating conditions is obtained, and the large-span arch bridge damage identification model is used to perform damage identification according to the actual response.
[0076] In a possible implementation, the memory 10 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and an application required for at least one function, etc.; the data storage area may store data created during use.
[0077] In addition, the memory 10 may include a read-only memory and a random access memory, and provide instructions and data to the processor. A part of the memory may also include NVRAM. The memory stores an operating system and operating instructions, executable modules or data structures, or a subset thereof, or an extended set thereof, wherein the operating instructions may include various operating instructions for implementing various operations. The operating system may include various system programs for implementing various basic tasks and processing hardware-based tasks. The processor 20 may be a central processing unit (CPU), a specific application integrated circuit, a digital signal processor, a field programmable gate array or other programmable logic device, and the processor 20 may be a microprocessor or may also be any conventional processor, etc. The processor 20 may call a program stored in the memory 10. The communication interface 31 may be an interface of a communication module for connecting to other devices or systems.
[0078] Of course, it should be noted that Figure 7 The structure shown does not constitute a limitation on the large-span arch bridge damage identification device in this embodiment. In actual applications, the large-span arch bridge damage identification device may include Figure 7 More or fewer components than shown, or combinations of certain components.
[0079] In an embodiment, a storage medium is further provided, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned large-span arch bridge damage identification method are implemented. The storage medium may include: a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.
[0080] In summary, the present invention makes up for the deficiency of large-span arch bridges due to the scarcity of historical monitoring data through transfer learning, fully extracts the information in the sparse labeled data under small samples compared to large-scale damage data sets, realizes the full use of structural health monitoring data, and improves the efficiency and robustness of damage identification of large-span arch bridges. At the same time, the damage index represented by the inner product matrix fully mines the damage information of the bridge structure, greatly improves the strong anti-noise ability in the field of damage identification of large-span arch bridges, and solves the problem that some vibration signals are changed under the influence of environmental interference, sensor instability and other factors, thereby affecting the detection results of bridge structure damage. Therefore, the present invention effectively overcomes the various shortcomings in the prior art and has a high industrial utilization value.
[0081] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. These modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention, and they should all be included in the scope of the claims and specification of the present invention.
Claims
1. A method for identifying damage of a long-span arch bridge, characterized in that: The long-span arch bridge damage identification method comprises the following steps: A bridge finite element model is constructed based on a long-span arch bridge, and structural acceleration data under various damage conditions are obtained using the bridge finite element model; An inner product matrix is established according to the structural acceleration data, and the inner product matrix is converted into an inner product matrix graph; Constructing a lightweight damage identification model, and preliminarily training the lightweight damage identification model using conventional training data; Using the inner product matrix diagram, fine-tuning the lightweight damage identification model after preliminary training is performed to obtain a large-span arch bridge damage identification model; The actual response of the long-span arch bridge under operating conditions is obtained, and the damage identification model of the long-span arch bridge is used to perform damage identification according to the actual response.
2. The damage identification method for a long-span arch bridge according to claim 1 is characterized in that: The method also includes performing noise adding processing on the structural acceleration data according to noise conditions.
3. The damage identification method for a long-span arch bridge according to claim 2 is characterized in that: The noise conditions include: A first noise operating condition, wherein the noise distribution of the first noise operating condition satisfies a normal distribution with a mean of 15 and a variance of 0; A second noise operating condition, wherein the noise distribution of the second noise operating condition satisfies a normal distribution with a mean of 15 and a variance of 1.0; A third noise operating condition, wherein the noise distribution of the third noise operating condition satisfies a normal distribution with a mean of 15 and a variance of 2.5; A fourth noise operating condition, wherein the noise distribution of the fourth noise operating condition satisfies a normal distribution with a mean of 15 and a variance of 5.
0.
4. The damage identification method for a long-span arch bridge according to claim 3 is characterized in that: According to the second noise condition and the third noise condition, the structural acceleration data is subjected to noise adding processing by using Latin hypercube sampling.
5. The long-span arch bridge damage identification method according to claim 1 is characterized in that: The step of establishing an inner product matrix according to the structural acceleration data comprises the following steps: Obtaining a time domain response of the structure according to the structural acceleration data; Setting a reference response, and obtaining an inner product vector according to the reference response and the time domain response; The inner product matrix is established through the inner product vector.
6. The long-span arch bridge damage identification method according to claim 5 is characterized in that: The inner product vector satisfies the following formula: R={R 1l (0),R 2l (0),…,R nl (0)} Where R represents the time domain response x of the lth measurement point. l (t) is the inner product vector of the reference response, R nl (0) represents the value of the cross-correlation function between the time domain response of the nth measuring point and the time domain response of the lth measuring point when the time delay τ=0.
7. The long-span arch bridge damage identification method according to claim 5 is characterized in that: The inner product matrix satisfies the following formula: R kl (0)=<x k (t),x l (t)>,k=1,2,…,n,l=1,2,…,n Among them, M represents the inner product matrix, R kl (0) represents the time domain response x of the kth measurement point k (t) and the time domain response x of the lth measurement point l is the value of the cross-correlation function of (t) when the time delay τ = 0, <·> represents the inner product operator, and n represents the number of measurement points.
8. The long-span arch bridge damage identification method according to claim 1 is characterized in that: The conventional training data is the ImageNet dataset.
9. The damage identification method for a long-span arch bridge according to claim 1 is characterized in that: The lightweight damage identification model is a Mobile ViT lightweight convolutional neural network.
10. A long-span arch bridge damage identification system, characterized in that: The large-span arch bridge damage identification system includes: an input device, an output device, a processor, and a memory. The input device, output device, processor, and memory are interconnected. The memory includes program instructions, and the program instructions are used to execute the large-span arch bridge damage identification method according to any one of claims 1 to 9.
